Looking behind occlusions: A study on amodal segmentation for robust on-tree apple fruit size estimation

Jordi Gené-Mola, Mar Ferrer-Ferrer, Eduard Gregorio-Lopez, Pieter M. Blok, Jochen Hemming, Josep Ramon Morros, Joan Ramón Rosell-Polo, Verónica Vilaplana, Javier Ruiz Hidalgo

Abstract: Highlights•A CNN for simultaneous modal and amodal instance segmentation was implemented.•Amodal segmentation was applied to predict visible and occluded apple regions.•Modal and amodal masks were used to estimate the % of visibility of apples.•The method was robust for detection and sizing of partially occluded fruits.•Results showed an F1 = 0.86 and a MAE = 2.93 mm for detection and sizing, respectively.